Fig. 13.

Relations between the number of epochs versus the f1-score for validation data and training loss for training data. The neural network training with the use of ’roberta-large_TDE’ architecture